By Debi Huffman, Ph.D.
Summary: Identifying effective indicators of virus removal in water treatment processes is an ongoing task. The following article strongly supports the case for bench-scale testing as opposed to stated membrane claims of microorganism removal. Moreover, we discover that bacteriophage may also play a key role in this procedure.
Recent advances in the development of low-pressure membranes along with their reduced cost have dramatically increased use of these membranes for drinking water treatment. Membrane removal capabilities are often based on pore size distribution or molecular weight cut-off (MWCO). MWCO is an approximate characterization of the membrane rejection capability, or the smallest compounds in terms of molecular weight that are rejected. These criteria, however, may not adequately describe the effective microbial removal capabilities of the membranes and, therefore, the level of public health protection provided.
Often, membrane descriptive classification such as micro-, nano-, or ultrafiltration, leave the water community with a number of unanswered questions regarding the membrane’s ability to remove microbial pathogens. This lack of understanding of membrane pore size may be compounded by the question of microbial size. What is smaller, bacteria or virus? Is the organism of interest a bacteria or a virus? To better answer these questions, the American Water Works Research Foundation (AWWARF) funded a recent research project under the direction of Montgomery Watson Harza (co-investigators from John Hopkins University, University of Colorado, University of South Florida, Thames Water and Vivendi Water) that strives to classify low-pressure membranes by their capability to remove one or more types of microorganisms under selected conditions using standardized protocols.
Study overseers
These protocols were reviewed by an advisory panel comprised of experts in the field of membrane processes as well as members of the U.S. Environmental Protection Agency (USEPA) and the National Sanitation Foundation, experts in the field of standards/protocol development for the testing and evaluation of water treatment technologies. Previous attempts to describe membrane effectiveness for microbial reduction have been performed at the bench-scale level. These results, however, were often difficult to compare due to the wide variability of testing protocols and equally variable microbial surrogate or pathogen selection. The new research seeks to compare the microbial removal obtained during bench-scale testing with that obtained using a similar membrane at either a pilot- or full-scale level.
Bench-scale testing of hollow fiber membranes was performed using a suite of microorganisms. Bacteriophages MS2 and PRD1 were chosen for use as surrogates for more infectious animal viruses. Bacteriophages are viruses that selectively infect only specific bacterial species. They’re quite harmless to humans and commonly used in microbial studies because of their ease of analysis (24 hours to results). They’re about 24 nanometers (nm) and 64 nm in size, respectively. Bacteriophage MS2 and PRD1 are often used as surrogates for polio- and rotavirus, which are 27 nm and 72 nm in size, respectively.
Viruses used in study
One animal virus was also utilized during bench-scale studies. The animal virus, the feline calicivirus (FCV), was used as a surrogate for the human infectious Norwalk-like virus, also known as the Norovirus. Feline calicivirus was selected for testing because it’s non-pathogenic (i.e., it doesn’t cause disease in humans) and is thus easier to use in the laboratory. Because this virus must be detected by growth in cell culture, however, it’s extremely time consuming and labor intensive to detect (days to weeks).
Bench-scale studies were also performed using strains of human enteric virus, poliovirus and hepatitis A virus (HAV), which is also 27 nm in size. These organisms can only be utilized in small-scale studies because they cannot be easily grown to high concentrations that would otherwise be necessary to use at the pilot- or full-scale level. Further, there’s a number of existing membrane bench-scale studies that have utilized these organisms in prior research, so comparisons can be made with previous studies.
Initial results of the bench-scale microbial removal studies have been evaluated for the correlation between microbial removal, MWCO (ranging from 10 to 300 kiloDaltons), pore size, and specific flux for different micro- or ultrafiltration membrane types. Two 10-kD (~5-nm pore size), four 100-kD (~14-nm), two 300-kD (~23-nm), and two 0.22-micron (µm) pore size membrane modules were challenged with the previously described microorganisms. (While we have approximated the pore size of the various membranes for ease of comparison, they are marketed based upon the MWCO that these filters are designed to achieve. The 0.22-µm membrane is marketed based upon pore size.) The average removal efficiencies of both bacteriophage and enteric viruses by the various membranes are shown in Table 1. 
As expected, 100- (~14-nm pore size) and 300-kD (~23-nm) membranes removed MS2, PRD1, poliovirus and FCV to the detection limits of their respective assays. The small pore-sized membrane, 10-kD (~5-nm pore size) manufacturer-rated membranes, however, only achieved significantly lower removals of MS2, PRD1, poliovirus and HAV. This was most unexpected, as the pore size of the 10-kD membrane is about four times smaller than the smallest challenge organism. The 0.22-mm membrane modules showed less than 0.2 log removals of MS2 and poliovirus. This was anticipated as both MS2 and poliovirus are of a size similar to the porosity of the membrane.
During bench-scale studies where the virus removal achieved the expected levels, the animal virus surrogate MS2 correlated (r2=0.96) with the removal of infectious animal viruses indicating the usefulness of MS2 as potential surrogate for membrane studies. These results also suggest that MWCO or hydraulic performance of membranes may not be indicative of microbial removal as indicated by 10-kD MWCO membranes having good specific flux but reduced microbial removal. Further bench and pilot-scale studies are being carried out with the results being presented at a number of national and international drinking water-related meetings including the American Water Works Association meeting last June, the International Water Association meeting in Cape Town, South Africa, in September, and the Water Quality Technology Conference in Pennsylvania, in November.
Conclusion
In short, bacteriophage has the potential to be a suitable indicator of infectious virus removal in water treatment processes using low-pressure membrane filtration. Membrane removal performance should be evaluated using bench-scale testing for microbial removal rather than relying solely on reported membrane characteristics for their abilities to remove potentially pathogenic microorganisms and protect public health.
About the author
Debi Huffman, Ph.D., is a research associate in the College of Marine Science at the University of South Florida and the director for the Center for Healthy Beaches/Healthy Coasts. She can be contacted at (727) 553-3946, (727) 553-1189 (fax) and email: dhuffman@marine. usf.edu.